Of people and computers

A few companies around the world are taking an entirely fresh look at the interface between people and computers, with surprising results

My father learned drafting not long before the U.S. entered World War II, and when he came home from the Pacific he began his peacetime career engineering industrial equipment, a job he does part time to this day. The firm he works for has all the usual 21st century computer-aided design (CAD) tools at their disposal, but Dad works on velum at a drafting board with an array of pencils and a drafting machine.

“What kinds of projects do they call you in to do now?” I asked him the other day.

The really tough ones,” he laughed. “I'm a lot faster with those kind of projects than CAD, although it works just fine for lots of things.”

This makes perfect sense to me. My father has built an astonishing “database” of knowledge and experience over the years and there are no algorithms restricting his random search capabilities. In fact, there are no constraints on his ability to find solutions at all.

We depend upon computing power so much today in the trucking industry and elsewhere, that it can be easy to forget the fact that people still significantly outperform even the most powerful computers at completing many tasks, particularly things like pattern recognition and making those imaginative leaps that delight corporate leaders who demand “creativity,” but are as mystified as the rest of us concerning its origins.

Now, however, a few companies around the world are taking an entirely fresh look at the interface between people and computers, with surprising results. Icosystem, for example, (www.icosystem.com) recently introduced a patented search and design technology dubbed “The Hunch Engine.” According to the company, it “is a way of doing [a] search when you don't know what you are looking for, but you'll know it when you find it.”

The Hunch Engine combines human intelligence with computer analysis to arrive at conclusions. The French Postal Service has already used it to help mail carriers create delivery routes that meet service requirements, but also permit workers to incorporate personal preferences like the opportunity to walk through a park along the way. The pharmaceutical industry is testing a version of the Hunch Engine to help chemists in their search for new drugs. Basically, the computer generates new ideas and combinations for the chemists to respond to. When they do, the computer takes that human input and goes to work again, guiding the users along in a “well, how about this, then?” sort of process.

Another Alberta-based enterprise called AskForCents (www.askforcents.com) has just completed what they describe as “alpha testing” on a computer system for sending requests to people (yes, people) who are willing to try to provide helpful replies in exchange for a small payment. The system was developed using a new software tool from Amazon.com called The Mechanical Turk (www.mturk.com).

AskforCents users can submit questions that require a factual answer or a subjective response. In other words, you can ask if there is train service from Dallas to Chicago or for recommendations on the best place to stay when you get there. “Although we try our best to provide helpful responses, we cannot guarantee the accuracy or value of the responses,” advises a notice on the website. “To make up for this, we try to supply responses from two different people for each request.”

While these projects are still in their infancy, we may look back one day to discover that they marked a crossroads, the point at which our single-minded devotion to all things digital was unexpectedly intersected once again by the rich and unique “analog” abilities of humans. If so, that will be a good day indeed.

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.